Among the older groups, most are white. However, among the younger groups, there is a very high proportion of Latinos. And the majority population in the 15 and younger age group are Latinos. If you project that out 10 to 20 years, it has clear consequences for the makeup of this vulnerable driving group.

Again, we have to attend to the population as it is at present. However, it is also critical to plan for the population as it will be.

What are the current implications of these population numbers for traffic fatalities?

For many decades, we have made progress in reducing the rate of fatal collisions by almost any measure. As measured per 100,000 licensed drivers, over time from 1992 to 2001, the most recent decade for which we have good data, there was a downward trend for all drivers until 2000.  Since then, as the trend for all drivers has started to creep up slowly, we see the trend for the drivers 19 to 20 and 21 to 25 rise much more steeply.

And this follows through in crashes where the driver has been drinking?

Yes. We have been making progress in reducing the rate of collisions with drinking involvement at all age groups and for all drivers. But we seem to be losing ground in recent years. And again, there are sharp increases for these target age groups.

How does this dovetail with what we already know about these younger drivers?

It is well-known that males have fatal collisions of all types at much higher rates than females. We still don't completely understand why. And the rate of fatal collisions peaks at these age groups, but what's interesting is that it drops for the older end of the age group and continues to fall rapidly.

Some questions I'd like to study are: Why is the rate so high for the younger age? Why does it drop so quickly? And is there any way to speed up this drop?

What do you see when you look at fatal crashes where the driver had been drinking?

In a substantial portion of the fatal crashes, the driver had been drinking. The rates by drivers' age parallel the rates for all fatal crashes, though of course they are lower. Males' rates are dramatically higher than females’. And again, the rate peaks at age 19 to 20 and then begins to drop. Once again, is there a way to accelerate this drop?

Is this trend borne out in the distribution of Primary Collision Factors? 

Yes. For our age groups, 19 to 20 and 21 to 25, DUI is the leading factor far and away, roughly twice as common as the next highest collision factor, which is excessive speed. Among the 31-to-35 age group, it has been cut by more than half and diminishes by more than half again among drivers who are older still.

And how do the data about ethnicity work into this analysis?

Most data from traffic crashes do not include references to ethnicity. But vital status data, which are collected by hospitals when they treat crash victims, include ethnicity.  These data tell us how the burden of traffic crashes does vary by ethnicity.

And on a population basis, as measured by the percentage of the population that they represent, not necessarily their percentage of licensed drivers, the group with the highest burden of mortality in traffic crashes is Latino youth.

How do these distinctions among these different populations play out in the DUI arena?

First, let me preface this by noting that DUI arrests declined for the past decade in California. It would be extremely interesting to know why this has been happening, especially considering the fact that the rates of collisions in which alcohol is involved have stopped falling and seem to even be on the rise.

How do age, gender and ethnicity play out in DUI arrests?

The gender differences continue to mirror what we saw earlier, with males experiencing far greater likelihood of being arrested than females. What's truly striking is the age element, especially among males. The arrest rate peaks for 21- to 25-year-olds at 3,000 (per 100,000 licensed drivers). That is roughly three times the rate for middle-aged drivers. Even more interesting, it drops dramatically at each of the next age groups, falling to about 2,200 for 26 to 30 and then just over 1,500 for 31 to 35.

These of course are consistent with the fatality rates we saw earlier. What is the reason for this? And the difference between males and females is truly striking: six times the arrest rate for males 21to 25 compared to females the same age. And what is the reason for this?

And what of ethnicity? 

The patterns show the usual age trend. But there is a very striking excess arrest rate for Latinos. And, again, when you look at the age groups we're talking about, 19 to 20 and 21 to 25, this is especially striking. Latinos in these groups are roughly 1.5 to 2 times as likely to be arrested as other ethnicities the same age. And they continue to lead all other ethnic groups until their mid to late 40s. And finally, the conviction rate is remarkably constant for age, and is fairly consistent across ethnic lines.

Can this be explained by bias in enforcement?

It could, though there's no evidence we know of. The arrest rates would appear to be noteworthy, given this same group's higher burden of mortality. Obviously, much more needs to be known, including possible enforcement and education interventions targeted to these groups.

Related Links:

Data presented at DUI Forum
(PowerPoint slide presentation-367.50KB)